Meta has abruptly shelved a major round of workforce cuts after its autonomous AI agents failed to deliver the promised productivity gains, according to internal documents reported by Reuters. Under a covert initiative codenamed "Project OT," CEO Mark Zuckerberg intended to slash several internal teams by up to 60 percent. The operational thesis rested on a lean, "talent-dense" skeleton crew supervising an army of virtual AI workers across daily pipelines.
That thesis collapsed on the evening of May 19, mere hours before the initial layoff notifications were set to deploy, when Zuckerberg pulled the plug on a second wave planned for November. The agent software underperformed in real operational workloads, while Wall Street began pushing back against the tech giant's escalating AI CapEx. By July, Zuckerberg openly conceded that agent capabilities had failed to mature at the pace leadership had projected.
The strategic blunder quickly sparked internal turmoil. Morale cratered after employees discovered background telemetry software logging their keystrokes and mouse clicks to train the very agents meant to replace them. The resulting internal uproar sent Meta's employee sentiment metric plunging from 74 percent to 55 percent, as remaining staff pushed back against babysitting brittle, hallucination-prone automated workflows.
For enterprise leadership eyeing aggressive headcount reduction through autonomous agents, Project OT offers an expensive cautionary tale: tying workforce cuts to premature automation milestones without proving operational reliability in actual pipelines produces managerial chaos and severe organizational friction rather than efficiency.